By Kenneth A. Shepsle and Mark S. Bonchek
This week’s political analysis book was “Analyzing Politics: Rationality, Behavior, and Institutions” by Kenneth A. Shepsle and Mark S. Bonchek. The book goes into detail about the rational choice model for analyzing politics. Many of the concepts in the book are things that I’ve learned in the context of past economics courses, and though some people refer to the rational choice model as the economic model of politics, in chapter one the authors are quick to dispute this belief and claim the model purely for politics. The book touches on a wide variety of topics, from models of individual choice, issues with groups choice, strategic behavior, cooperation and public goods, on to a discussion of a variety of institutions, both U.S. and foreign.
Part I: Introduction
Part one begins by explaining the origins of rational choice theory. The authors note that before WWII, political analysis was based on detailed descriptions of events and judgments about what should have been done. Since WWII, political analysis has evolved to focus on explanation and analysis of events – looking at which items are particularly important and trying to understand why things happened the way they did. It defines the model of rational choice as one in which “individuals act in accordance with their preferences for outcomes and their beliefs about the effectiveness of various actions available to them.” In other words, people use their understanding of the world and beliefs about the consequences of their actions to make decisions that will result in achieving their preferences and maximizing their utility. Of course, people aren’t always sure what actions will actually result in the outcomes they prefer, and to take into account this uncertainty, rational choice uses the concept of expected utility – essentially the product of the uncertainty and utility associated with the possible outcomes of a particular choice.
Part II: Group Choice
In Part II, the focus moves from individual choice to group choice. The major finding is that even if individuals honestly reveal their preferences, it is possible for group preferences to be incoherent – if the procedures for voting (simple plurality, round-robin, run-off) are changed, then the resulting group preference that is expressed can be different. (Normally, with an individual, if person A prefers A to B and prefers B to C, then he also prefers A to C. When groups vote, this type of rule doesn’t always hold – that’s what make the result incoherent.) This situation allows voting to result in Condorcet’s Paradox, in which voting is cyclical, and any currently winning group choice could be beat if pitted against another possible outcome. The problem of incoherent group choice is formally defined in Arrow’s Theorem, which states, “There exists no mechanism for translating the preferences of rational individuals into a coherent group preference that simultaneously satisfies basic conditions for fairly making group decisions.” Essentially, there is no way to design a voting procedure to be sure you have a coherent group preference result without resulting to tyranny (i.e. everyone defers to one leader) or some other unfair method. This means that there is no method of majority rule that can guarantee coherent group choice – so much for democratic voting!
The section continues by giving some exceptions. One is Black’s Single-Peakedness Theorem, which basically says that if there is an alternative that no individual in the group ranks as worst, then majority rule can generate transitive (coherent) group preferences. Sen extends this to say that as long as there is one item that no individual chooses to rank top, middle, or bottom, then there can be coherent group preferences. For example, Black developed his theory by noticing that in Northern Ireland the Protestants preferred an agreement linking them to England to the Status Quo, but preferred the status quo to having Northern Ireland become part of the Republic of Ireland. Catholics preferred becoming part of the Republic of Ireland to the Status Quo, but preferred the Status Quo to an agreement linking them to England. In this case, a decision to stick with the Status Quo would be a stable equilibrium – no majority would want to move from that point. The authors show methods of creating special models to show graphically this “single-peakedness” condition. This is used to show that the median choice is the stable solution. At the median preference, there is no majority that would prefer to move the decision to another spot. However, moving to multi-dimentional spatial analysis, (except under very rare conditions where alternating preferences are symmetric) they show that the equilibrium is lost and Condorcet’s cycling can occur again.
The one-dimensional analysis helps to explain why in two-candidate races it is common for the candidates to position themselves towards the middle of the political spectrum. If the political spectrum is imagined as a line, then a candidate wants to be as close to the middle as possible, while not crossing past the other candidate –this allowing them to capture all the votes to the left of middle (if they’re liberal), since even people on the far end (very liberal) prefer someone slightly left of center to someone slightly right of center. It can also be shown why if a third candidate can enter a race, candidates will be farther from the center. In the case where a third candidate could enter, the two main candidates would want to be sure there weren’t sufficient votes between them for a moderate to come and take the moderate votes, but they also want to be sure that they aren’t so far from the edge of the line that a third party candidate could come in slightly to the left of them (for a liberal candidate) and take a majority of the more extreme candidates.
Two Candidates (no entry allowed): L----------lr----------R – Candidates place themselves near center to capture most possible votes
Two Candidates (entry allowed): L------l------r------R – Candidates arrange themselves so that a third candidate cannot place themselves on the line in a way that results in the third candidate getting a plurality of the votes.
Median voting can also be shown to apply to legislative voting – again causing the median preference to be the stable solution. Also, where bills are written in committees, the committee median preference will be the agreed-upon bill. Then if a bill is brought from committee to the house or senate under closed rules (no voting allowed), then Congress will have choose the committee median, as long as it is closer to the Congress median than the Status Quo. Under open rules, the committee bill will be amended until it is the same as the Congress median preference.
Another issue dealing with group choice is the opportunity for individuals to use strategic behavior. In undertaking strategic or sophisticated voting, an individual uses rational foresight – i.e. thinks about what is likely to happen in the future – and uses this knowledge to possibly vote against their own preferences to get a preferred outcome. Sophisticated voting includes thinking ahead about how others are going to vote in the future to decide how you should vote – this method is called backwards induction in game theory. This type of logic is used when Congressmembers make their decisions about voting based on an understanding of the chances the president will veto a particular bill – i.e a Congressmember may dislike a popular bill, but vote for it anyway, knowing that the president has already decided to veto. Strategic voting doesn’t involve working backwards from other people’s actions, but rather is based on changing your vote to minimize the chances of getting either your most- and least-preferred outcome. This is the type of thinking that occurs when individuals whose first choice In an election is a third party candidate recognize that their preferred candidate is unlikely to win and instead vote for their second choice candidate, to help minimize the chances that their least-liked candidate would win. (Like a Nader fan voting for Gore to avoid having Bush.) This kind of strategic activity can also take place on a larger level in Congress – a policy entrepreneur may change the framework with which an issue is discussed to hide their true motives and gain more support.
It was previously mentioned in section one that voting methods matter, and the last part of section two expands on what this means for the electoral system. It details lots of types of electoral systems, including: simple plurality (like U.S. elections), plurality run-off, ranking all preferences, allowing multiple votes per person, and a number of others. In general, Duverger’s law shows that electoral systems based on simple plurality, like the U.S. system, tend to be associated with two-party rule, and that proportional representation, where representatives are given seats proportional to the percentage of the national vote that they won, are associated with multiple parties.
Part III: Cooperation, Collective Action, and Public Goods
Part III of the book covers cooperation, collective action, and public goods. It starts out by discussing simple two-person cooperation using the Prisoner’s Dilemma example from game theory. Though the equilibrium solution in the Prisoner’s Dilemma is for each participant to choose not to cooperate, there are methods in which cooperation can be reached. One was identified by Axelrod, who suggested that if the game is infinitely repeated (you have lots of opportunities for cooperation with the same person over time) then the ‘shadow of the future’ (i.e. the knowledge that you’ll probably have more opportunities to cooperate) makes you more likely to cooperate now. He also suggests the best strategy is a ‘tit for tat’ strategy, in which you begin by cooperating, but then copy whatever the other player did in the previous game. Another method for getting cooperation is actually changing the game’s pay-out with internalized values. For example, if the two criminals in the prisoners dilemma were taught that it isn’t right to squeal, they may value staying quiet higher than would be expected based on economic benefit. Similarly, external enforcement can change the payout – if you know that snitching will result in punishment from some outside force (the mafia boss) – choosing to stay quiet may be more valuable than it otherwise seems. In society, we try to enforce cooperation through laws, contracts, and other things, but it is difficult because enforcement is costly and imperfect – you might not want to pay the fees to bring someone to court and you can’t be sure (even if you’re ‘right’) that the judge will rule in your favor.
The issue of collective action takes cooperation to the next level. In groups it’s difficult to get people to cooperate because it is possible for them to be a ‘free rider’ – i.e. to benefit from everyone elses actions without actually contributing. One theory suggests that its easier to get participation if a person believes their participation is essential to achieving the goal. Mancer Olson suggests that people join groups because of the by-products available only to members – information, discounts, materials, etc. Some groups may form because political entrepreneurs identify a latent group and help to organize them – for example, a Congress member might help to pass laws to make it easier for labor unions to form in a particular area. It’s also possible that some people join groups just because they enjoy the group or believe in its cause, though evidence suggests that this is done more often for non-economic groups, like environmental groups, than for groups such as labor unions.
Applying the issue of collective action to voting seems to makes voting in national elections seem irrational – the impact of one vote and personal benefit of voting are unlikely to outweigh the costs of getting to the voting booth. The book suggests that people vote not because of this cost-benefit trade-off, but rather because they enjoy it, or other factors (like social pressure) encourage them to do so.
The final chapter in Part III discusses public goods, externalities, and the commons. Public goods are those defined as non-excludable – you can’t stop others from using/enjoying it, and non-rivalrous – just because one person is using it doesn’t mean that another person can’t use it – it isn’t “used up”. Examples include lighthouses, the ocean, or national security. The private market undersupplies these items, since there is no way to force users to pay for them (since they’re non-excludable). This means that governments are often the suppliers of public goods. However, for a good to be provided, it is likely necessary for someone to push the government to do so. Interestingly, it is often not the consumers of these goods that push governments to provide them, but rather the private producers. For example, it is often the concrete industry, not drivers, that is the main group pushing for new and better roads. In cases where the government does not pay for private producers, like in the area of basic science research, the provision of goods is often ineffiecient because of the political process. Science funding is sent to many different areas of the United States to satisfy Congressmen rather than to only those areas with the most cutting-edge labs or best qualified scientists.
Externalities and Commons are examples of other items with public goods properties. Externalities are the extra goods (or bads) that are produced as a by-product of activity and are not paid for by the producer – pollution is a classic example. The government can choose to remedy the production of externalities with taxes (on ‘bads’), subsidies (on ‘goods’), or through privatization schemes – like cap and trade on pollution emissions. Issues of the commons – parks, fishing areas, etc. – is similar to that of other public goods and externalities. To avoid over-fishing, the government likely needs to implement a scheme to provide permits or monitor fisherman to limit the size of the catch.
Part IV: Institutions
The final section of the book was about institutions. In general, it noted that institutions all allow the division of labor and focus on particular issues (jurisdictions), which allows the creation of experts in particular areas. Institutions also develop procedures and rules to help them perform repeated tasks more efficiently. Within each institution discussed, the book examines how rational choice can be used to understand how that institutions comes to the decisions it does.
The first institution that is looked at is the legislature, and how individual legislators make decisions. The book notes that though all legislators are accountable to constituents, but that some see themselves as ‘delegates,’ there simply to represent their constituents interests, while others believe they are ‘trustees,’ chosen to go to Congress and do what they believe is the ‘right’ thing. Either way, it is important to note that legislatures have heterogeneous (different) preferences.
Going back to earlier discussions about cooperation, it is possible to apply the issues of group decision making to voting within Congress. Since many issues are multi-dimensional, it is possible to get cycles in voting, and incoherent preferences. Another interesting aspect of legislators decision-making is their greater interest in some areas than others. Representatives of coastal areas may be very interested in maritime law, while a representative from Wyoming is not. However, Congressmemeber only have one vote per bill, regardless of the issue. Also complicating this is the fact that they only have limited information on what instrumental actions will actually lead to their preferred outcomes (i.e. which policy will actually ‘work’), and the fact that whatever choice they make needs to be carried out by the bureaucracy which may or may not comply exactly with their choice.
One coping mechanism has been for Congressmen to split up into committees. This allows them to focus on the area in which they are particularly interested. They also become experts in this area and are better able to judge the likely effect of various policy actions. Committees are also in charge of monitoring compliance of the bureaucracy. Some committees are very specialized, such as agriculture, fishing, or others, and these committees often are not representative of Congress overall. This is important, because as mentioned in section two, under closed rule, a committee can impose its particular preferences on the group (within some bounds). However, since these issues are so specialized, this often isn’t an issue for members of the wider group. There are also more general committees, such as the Finance Committee or the Foreign Affairs Committee, and membership for these committees is chosen to be more representative of the Congress as a whole, avoiding the important and wide-ranging issues from being controlled by a non-representative group.
The next institution examined was the bureaucracy and intergovernmental relations. The book offered three theories on how the bureaucracy functions. The first theory sees the bureaucracy as budget-maximizing, and the legislature as a passive sponsor. The bureaucrats use information about their own costs and the legislature’s demand to maximize their own budget and get the most ‘on-the-job consumption’ (nice offices, trips, etc.) or at least prestige, that’s possible. This results in over production of the bureaucratic good. The second theory suggests that bureaucrats and legislators have some knowledge about each others’ cost and demand, and use this to bargain over the correct budget level, but that uncertainty can result in over production or under production of the bureaucratic good. The third theory suggest that the legislature is the principle and the bureaucracy their agent, so that Congress can monitor what the bureaucracy does. Though there may still be some bureaucratic drift (i.e. the bureaucracy doesn’t do exactly as they’re told), this results in closer alliance between legislative demand and bureaucratic supply. In addition to bureaucratic drift, it is also possible for the oversight committee and the bureaucracy to ‘drift together,’ and this situation will ‘snap back’ to the median if the representation on the legislative oversight committee changes.
Rather than discuss the presidency, prime minister position, or some other specific leader, the book discusses leadership in general. In general, the leader is representing a group and needs to have the support of that group. As long as the leader maintains support, he/she usually has the power to set the agenda (which we saw previously can affect the outcome of a vote). Even if not asked, a person may become a leader by bringing together a latent group, as was discussed previously with regard to collective action. In any case, it’s important for a leader to gain a reputation – this may be done by showing a willingness to punish early in his/her tenure to gain compliance without having to actually punish anyone later on, thus retaining the support of the group.
Courts and Judges represented an interesting case for rational choice theory, since judges and supreme court justices do not run for election (they serve life terms) and they don’t have performance pay. It is suggested that judges are instead motivated by the prospect of career-enhancement, prestige, or a general enjoyment of the job (so the job is more like consumption than production). The book also suggests that judges may be thought of as ‘legislators in robes’ who are motivated by the ability to influence policy. It’s impossible for judges to avoid incorporating their own values to some extent. Even if judges don’t often ‘make policy,’ legislators and bureaucrats take into account the role of the court in deciding how extreme the policies they make and implement should be, so the threat of the court over-turning a policy may be enough to affect policy.
The final chapter in the book applies rational choice theory to governments outside the United States. They show that electoral arrangements have a large affect on voting outcomes, as discussed in part two of the book. In simple plurality elections, compromises are made before the election, and there are often only two parties. In proportional representation, many parties may be elected, and coalitions are likely to be formed after election.
In places with proportional elections, the legislatures elects the ‘government’ – the prime minister and ministers of various departments. Based on the various groups within legislatures, the make-up of the government can vary. The incidence of different types of governments since 1945 is: Single party majority rule (.134), multiple party majority (.5), single party minority (.238), multiple party minority (.128). The book suggests that the choice of how to form a government may be affected by office-seeking parties, who wish to maximize the number of offices to which its members are assigned. Another possibility is based on policy-seeking groups who aim to create a coalition as close as possible to their ideal policies. The book lays out these decisions using spatial modeling, similar to that used in section two.
Overall, the idea of this book was to layout the various issues associated with the rational choice theory – considerations when trying to understand group choice, difficulties of collective actions, and the provision of public goods. The book ends by applying these various considerations and methods to political institutions both within the U.S. and without.